Equating LLMs and transformers is not a meaningless terminology difference at all, Aristotle is so different from the things people call LLMs in terms of training data, loss function, and training that this is a grievous error.